[justinchuby]

Summary

• Corrective follow-up to already-merged docs: publish wiki with Quartz 5 (replaces #1210) #1272; applies the finalized maintainability-first Quartz tree without replaying the obsolete branch history.
• Replaces the custom Acorn AST/dataflow/resource-graph audit and local graph vendoring with a 54-line deterministic runtime pin and an exact four-URL immutable allowlist.
• Retains source wikilink, repository-target, generated link/asset, base-path, referenced-resource, and rendered-root checks.
• Publishes wiki/index.md as the real site landing page while keeping wiki/README.md as the conventions/index note.
• Removes direct Acorn, D3, and Pixi dependencies while preserving the official Pages workflow and pinned Quartz/plugin/action revisions.

This simplifies #1272; it does not reopen #1210.

Validation

• npm ci
• Clean CI=true npm run wiki:build, followed by a consecutive production build
• 18 focused tests passed
• 116 wikilinks validated across 21 notes
• 90 repository source links rewritten from verified repository targets
• 3,198 generated internal links/assets validated across 118 HTML pages; 2 referenced local scripts verified
• 42 pinned Quartz plugin commits/build outputs verified
• Exact immutable runtime allowlist: 4 URLs
• 127 generated files matched byte-for-byte across consecutive builds
• npm run check (TypeScript + Prettier), Python compile checks, YAML/JSON parsing, and git diff --check
• Custom audit/validation code: 2,875 -> 579 lines (2,296 fewer; 79% reduction)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 80.79%. Comparing base (3edaa16) to head (9079cb1).
⚠️ Report is 43 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##           main    #1280       +/-   ##
=========================================
+ Coverage      0   80.79%   +80.79%     
=========================================
  Files         0      364      +364     
  Lines         0   160710   +160710     
  Branches      0   160710   +160710     
=========================================
+ Hits          0   129840   +129840     
- Misses        0    26214    +26214     
- Partials      0     4656     +4656     

Flag	Coverage Δ
mlas	85.32% <ø> (?)
offline	80.70% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.
see 364 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

🔴 Benchmark Regression Detected

Comparison of criterion micro-benchmarks: PR head vs merge-base, measured on the same runner in the same job (base first → PR second).

ℹ️ Absolute times are informational only — they vary with runner load. The % change column is the reliable signal because both sides ran under identical conditions.

Status	Scenario	Base	PR	Change
🔴	gather/large_f32_threads=1-internal/131072	28.19 µs	47.44 µs	+68.2%
🔴	gather/large_bf16_threads=1-internal/131072	10.25 µs	15.42 µs	+50.5%
🔴	matmul/medium_generic_bf16_threads=8/32x512x512	356.08 µs	483.34 µs	+35.7%
🔴	block_quantized_matmul_cached_dense/mxfp4_cached_dense_repeated_call/1x1024x1024	48.38 µs	65.45 µs	+35.3%
🔴	block_quantized_moe_cached_dense/mxfp4_cached_dense_expert_repeated_call/rows=1,H=256,I=256,E=4,top_k=1	54.23 µs	71.84 µs	+32.5%
🔴	gather/large_f16_threads=1-internal/131072	10.82 µs	14.30 µs	+32.1%
⚠️	block_quantized_matmul_cached_dense/mxfp4_preexpanded_dense_oncelock_like_proxy/1x1024x1024	44.70 µs	53.97 µs	+20.7%
⚠️	matmul/medium_generic_f16_threads=8/32x512x512	29.24 µs	35.16 µs	+20.2%
⚠️	block_quantized_matmul_cached_dense/mxfp4_uncached_dequant_each_call/1x1024x1024	655.35 µs	770.70 µs	+17.6%
⚠️	matmul/medium_generic_bf16_threads=1/32x512x512	533.27 µs	621.89 µs	+16.6%
✅	matmul/small_generic_f32_threads=8/1x256x256	36.51 µs	41.82 µs	+14.5%
✅	reduce_mean/large_f32_threads=1-internal/262144	999.59 µs	1.14 ms	+14.0%
✅	matmul/medium_generic_f16_threads=1/32x512x512	30.27 µs	34.01 µs	+12.4%
✅	gather/small_f16_threads=1-internal/4096	529.9 ns	595.2 ns	+12.3%
✅	add/large_f32_threads=1-internal/4194304	761.99 µs	851.66 µs	+11.8%
✅	matmul/small_generic_f16_threads=8/1x256x256	30.44 µs	33.80 µs	+11.0%
✅	matmul/large_generic_f32_threads=8/32x1024x1024	3.95 ms	4.35 ms	+10.2%
✅	block_quantized_moe_cached_dense/mxfp4_uncached_expert_dequant_each_call/rows=1,H=256,I=256,E=4,top_k=1	383.91 µs	419.76 µs	+9.3%
✅	matmul/large_generic_bf16_threads=8/32x1024x1024	1.34 ms	1.46 ms	+9.2%
✅	gather/small_f32_threads=1-internal/4096	688.8 ns	748.3 ns	+8.6%
✅	matmul/medium_generic_f32_threads=1/32x512x512	2.30 ms	2.47 ms	+7.4%
✅	add/medium_bf16_threads=1-internal/262144	111.92 µs	119.97 µs	+7.2%
✅	gather/medium_f16_threads=1-internal/32768	2.46 µs	2.61 µs	+6.4%
✅	matmul/medium_generic_f32_threads=8/32x512x512	975.71 µs	1.03 ms	+6.0%
✅	matmul/small_generic_bf16_threads=1/1x256x256	31.65 µs	33.46 µs	+5.7%
✅	matmul/small_generic_bf16_threads=8/1x256x256	30.94 µs	32.36 µs	+4.6%
✅	matmul/small_generic_f16_threads=1/1x256x256	31.37 µs	32.62 µs	+4.0%
✅	matmul/large_generic_bf16_threads=1/32x1024x1024	2.03 ms	2.09 ms	+2.9%
✅	matmul/large_generic_f16_threads=1/32x1024x1024	79.89 µs	82.08 µs	+2.7%
✅	gather/medium_bf16_threads=1-internal/32768	2.47 µs	2.53 µs	+2.3%
✅	add/small_bf16_threads=1-internal/1024	482.6 ns	487.6 ns	+1.0%
✅	reduce_mean/medium_f32_threads=1-internal/65536	266.85 µs	268.18 µs	+0.5%
✅	matmul/small_generic_f32_threads=1/1x256x256	40.00 µs	39.95 µs	-0.1%
✅	gather/medium_f32_threads=1-internal/32768	4.60 µs	4.60 µs	-0.1%
✅	sampling_latency/top_p_per_token	396.16 µs	392.86 µs	-0.8%
✅	add/medium_f32_threads=1-internal/262144	28.78 µs	28.37 µs	-1.4%
✅	logit_processing/seven_processor_chain_per_step	335.22 µs	327.52 µs	-2.3%
✅	gather/small_bf16_threads=1-internal/4096	522.1 ns	506.1 ns	-3.1%
✅	matmul/large_generic_f32_threads=1/32x1024x1024	10.07 ms	9.73 ms	-3.4%
✅	add/small_f16_threads=1-internal/1024	511.4 ns	492.4 ns	-3.7%
✅	matmul/large_generic_f16_threads=8/32x1024x1024	91.43 µs	87.91 µs	-3.8%
✅	sampling_latency/min_p_per_token	230.00 µs	217.87 µs	-5.3%
✅	add/small_f32_threads=1-internal/1024	237.8 ns	224.0 ns	-5.8%
✅	qwen3_sampling_processors/top_k_partial_selection	156.34 µs	146.95 µs	-6.0%
✅	qwen3_sampling_processors/top_k_full_sort_baseline	2.35 ms	2.20 ms	-6.3%
✅	qwen3_sampling_processors/top_p_full_sort_after_top_k_baseline	3.91 ms	3.61 ms	-7.5%
✅	add/large_bf16_threads=1-internal/4194304	1.88 ms	1.74 ms	-7.7%
✅	sampling_latency/top_k_per_token	58.03 µs	52.90 µs	-8.8%
✅	kv_cache/alloc_dealloc_pages	42.72 µs	38.82 µs	-9.1%
✅	qwen3_sampling_processors/top_p_fast_after_top_k	587.39 µs	533.70 µs	-9.1%
✅	add/medium_f16_threads=1-internal/262144	126.19 µs	113.76 µs	-9.9%
✅	qwen3_sampling_processors/top_k_top_p_fast	755.64 µs	669.37 µs	-11.4%
✅	qwen3_sampling_processors/top_k_top_p_full_sort_baseline	6.59 ms	5.83 ms	-11.6%
✅	grammar_masking/llguidance_compute_mask/32	90.45 µs	78.99 µs	-12.7%
✅	sampling_latency/greedy_per_token	3.77 µs	3.24 µs	-14.0%
✅	tokenization/decode_tokens_per_second	7.49 ms	6.44 ms	-14.1%
✅	tokenization/encode_tokens_per_second	456.79 µs	389.13 µs	-14.8%
🟢	reduce_mean/small_f32_threads=1-internal/4096	18.69 µs	15.56 µs	-16.8%
🟢	add/large_f16_threads=1-internal/4194304	2.33 ms	1.73 ms	-26.0%

Visual flags: ⚠️ ≥ 15% slower, 🔴 ≥ 30% slower — calibrated against measured runner noise (~27% worst-case on multi-threaded matmul)

Host info

CPU: Apple M1 (Virtual)
Cores: 3
OS: Darwin 25.5.0 arm64
Rust: rustc 1.97.1 (8bab26f4f 2026-07-14)
Load avg: { 3.81 3.38 6.86 }

What this cannot catch

• Regressions in code paths not covered by these benchmarks (e.g., end-to-end decode with a real model)
• Sub-threshold regressions that compound over multiple PRs
• Performance changes that only manifest under GPU execution
• Latency changes in the ORT integration path (these benchmarks exercise the native Rust kernels)